#include <qbenchmark_p.h>
Definition at line 123 of file qbenchmark_p.h.
◆ QBenchmarkTestMethodData()
| QBenchmarkTestMethodData::QBenchmarkTestMethodData |
( |
| ) |
|
|
default |
◆ ~QBenchmarkTestMethodData()
| QBenchmarkTestMethodData::~QBenchmarkTestMethodData |
( |
| ) |
|
◆ adjustIterationCount()
| int QBenchmarkTestMethodData::adjustIterationCount |
( |
int | suggestion | ) |
|
◆ beginDataRun()
| void QBenchmarkTestMethodData::beginDataRun |
( |
| ) |
|
◆ endDataRun()
| void QBenchmarkTestMethodData::endDataRun |
( |
| ) |
|
◆ isBenchmark()
| bool QBenchmarkTestMethodData::isBenchmark |
( |
| ) |
const |
|
inline |
◆ resultsAccepted()
| bool QBenchmarkTestMethodData::resultsAccepted |
( |
| ) |
const |
|
inline |
◆ setBlockSize()
| void QBenchmarkTestMethodData::setBlockSize |
( |
int | split = 32 | ) |
|
Split up iterations into a statistically useful number of blocks.
Record data in blocks. split is the desired number of blocks which is a power of 2 like iterationcount. As long as split is below 32 we treat each iteration as a block. When split is greater than 32 split becomes a divisor for iterationcount. The goal is to reduce overhead from measurements while calculating variance. The default split 32 should be sufficient to get stable block-variance, from which the per-iteration variance (estimator) can be computed by dividing by the block size.
Definition at line 175 of file qbenchmark.cpp.
◆ setResult()
◆ setResults()
◆ current
◆ iterationCount
| int QBenchmarkTestMethodData::iterationCount = -1 |
◆ measurementBlockSize
| int QBenchmarkTestMethodData::measurementBlockSize = 1 |
◆ resultAccepted
| bool QBenchmarkTestMethodData::resultAccepted = false |
◆ results
◆ runOnce
| bool QBenchmarkTestMethodData::runOnce = false |
◆ valid
| bool QBenchmarkTestMethodData::valid = false |
The documentation for this class was generated from the following files: